Identifying Traffic Features of Braess Links in Road Networks

Abstract

Braess links affect the performance of a road network. It is very important to detect and identify Braess links for improving the efficiency of the transportation network. Existing methods identifying Braess links include: traffic distribution, structure identification and traffic features identification. The first two methods have issues with bad maneuverability, while the traffic feature in the third method was often chosen inappropriately. This study addresses these issues by conducting experiments to detect the traffic features of Braess links in an experimental setting that resembles a real road network with multi- origin and destination (OD) pairs, non-linear congestion costs and two-way links. The experimental process includes eliminating links, taking U-E and O-S traffic distributions and varying demands. The minimum saturation rate was found as a notable feature for a Braess link. The cause of this phenomenon was explained. The study provides a practical way for optimizing the design of road networks.